SportingIntellect said:
Common Steve. That was a bit dramatic the trend didn't have a ton of filters, now I'm with you on this one not having all that much predictive value though. But I do think Ed showing off the things the SDQL can do is a good thing for people looking to learn to do them, themselves hope you both have great seasons.
I appreciate the clarification SI. I agree that the two trends Steve is equating:
"The 49ers are 12-0 ATS when they are off a road game in which Frank Gore did not have a reception of 10-plus yards." and
"Under a Harvest Moon on an odd Day in October, provided both games' QBs throw for an odd amount, AND the Bengels Tends catch UNDER 3 passes, and their opponents score between 17-24 but not 20 or 23 the Bengals are 44-0 ATS."
are quite different and do not warrant his ridicule.
Also, thanks for pointing out that I'm not recommending that anyone blindly follow trends. I'm simply demonstrating the depth and breadth to which historical results in the NFL can be queried with the SDQL -- fast and free.
In Steve's defense, I don't think he was addressing the comment to me, he was simply warning the members of the forum that blinding playing trends is not likely to be profitable -- especially ones that are over-fitted, as his example clearly demonstrates. Steve and I are in complete agreement here and I don't mind the reminder.
I think we are ALL in agreement that more information is better. The key to handicapping success is to quickly discard the information that is simply a random fluctuation and makes no handicapping sense and to cherish the information that makes good handicapping sense, is timely and has nice average margins.
Using the SDQL is not a handicapping STYLE, it is simply a tool to improve any style of handicapping.
Dr M.